Increasingly Packing Multiple Facial-Informatics Modules in A Unified Deep-Learning Model via Lifelong Learning

Benchmark Model Rank Results
age-and-gender-classification-on-adiencePAENet (single crop, tensorflow)#5Accuracy (5-fold): 89.08
age-and-gender-classification-on-adience-agePAENet (single crop, tensorflow)#6Accuracy (5-fold): 57.3
continual-learning-on-cifar100-20-tasksPAENet#6Average Accuracy: 77.1
facial-expression-recognition-on-affectnetPAENet#26Accuracy (7 emotion): 65.29